Lippmann R P, Shahian D M
Department of Thoracic and Cardiovascular Surgery, Lahey Hitchcock Medical Center, Burlington, Massachusetts 01805, USA.
Ann Thorac Surg. 1997 Jun;63(6):1635-43. doi: 10.1016/s0003-4975(97)00225-7.
Neural networks are nonparametric, robust, pattern recognition techniques that can be used to model complex relationships.
The applicability of multilayer perceptron neural networks (MLP) to coronary artery bypass grafting risk prediction was assessed using The Society of Thoracic Surgeons database of 80,606 patients who underwent coronary artery bypass grafting in 1993. The results of traditional logistic regression and Bayesian analysis were compared with single-layer (no hidden layer), two-layer (one hidden layer), and three-layer (two hidden layer) MLP neural networks. These networks were trained using stochastic gradient descent with early stopping. All prediction models used the same variables and were evaluated by training on 40,480 patients and cross-validation testing on a separate group of 40,126 patients. Techniques were also developed to calculate effective odds ratios for MLP networks and to generate confidence intervals for MLP risk predictions using an auxiliary "confidence MLP."
Receiver operating characteristic curve areas for predicting mortality were approximately 76% for all classifiers, including neural networks. Calibration (accuracy of posterior probability prediction) was slightly better with a two-member committee classifier that averaged the outputs of a MLP network and a logistic regression model. Unlike the individual methods, the committee classifier did not overestimate or underestimate risk for high-risk patients.
A committee classifier combining the best neural network and logistic regression provided the best model calibration, but the receiver operating characteristic curve area was only 76% irrespective of which predictive model was used.
神经网络是一种非参数、稳健的模式识别技术，可用于对复杂关系进行建模。
利用胸外科医师协会1993年80606例接受冠状动脉旁路移植术患者的数据库，评估多层感知器神经网络（MLP）在冠状动脉旁路移植术风险预测中的适用性。将传统逻辑回归和贝叶斯分析的结果与单层（无隐藏层）、两层（一个隐藏层）和三层（两个隐藏层）MLP神经网络进行比较。这些网络采用随机梯度下降法并结合早期停止进行训练。所有预测模型使用相同的变量，并通过对40480例患者进行训练以及对另外40126例患者进行交叉验证测试来评估。还开发了一些技术来计算MLP网络的有效比值比，并使用辅助“置信MLP”为MLP风险预测生成置信区间。
包括神经网络在内的所有分类器预测死亡率的受试者工作特征曲线面积约为76%。使用一个由MLP网络和逻辑回归模型输出平均得到的双成员委员会分类器时，校准（后验概率预测的准确性）略好。与单个方法不同，委员会分类器不会高估或低估高危患者的风险。
结合最佳神经网络和逻辑回归的委员会分类器提供了最佳的模型校准，但无论使用哪种预测模型，受试者工作特征曲线面积仅为76%。
